A practical one-pot radical-ionic sequence for the preparation of epoxides: Application to the synthesis of unnatural polyhydroxylated alkaloids

Article abstract of DOI:10.1016/j.tetlet.2011.10.042

An efficient one-pot sequence for the preparation of epoxides from α-iodoesters or α-iodonitriles and allylic alcohols is described. This sequence is based on the use of iodine atom transfer reaction onto allylic alcohols followed by a ring closing epoxidation reaction of the halohydrin intermediates. The feasibility of this sequence is showcased in the synthesis of the perhydroaza-azulene, an unnatural analog of castanospermine.

Full text of DOI:10.1016/j.tetlet.2011.10.042

Tetrahedron Letters 52 (2011) 6899–6902
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
A practical one-pot radical-ionic sequence for the preparation of epoxides:
application to the synthesis of unnatural polyhydroxylated alkaloids
⇑
Eduardo Peralta-Hernández, Omar Cortezano-Arellano, Alejandro Cordero-Vargas
Instituto de Química, Universidad Nacional Autónoma de México, Circuito Exterior s/n, Ciudad Universitaria, 04510 México D.F., Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
An efficient one-pot sequence for the preparation of epoxides from
a-iodoesters or a-iodonitriles and
Received 20 September 2011
Revised 5 October 2011
Accepted 7 October 2011
Available online 15 October 2011
allylic alcohols is described. This sequence is based on the use of iodine atom transfer reaction onto allylic
alcohols followed by a ring closing epoxidation reaction of the halohydrin intermediates. The feasibility of
this sequence is showcased in the synthesis of the perhydroaza-azulene, an unnatural analog of
castanospermine.
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Free radical
Atom transfer reaction
Epoxides
Polyhydroxylated alkaloids
OH
OH
HO
HO
HO
H
N
H
N
H
N
OH
OH
Introduction
HO
HO
Since their isolation, polyhydroxylated indolizidine alkaloids
like (+)-castanospermine (1),1 (ꢀ)-swainsonine (2),2 and (+)-lentig-
inosine (3)3 have attracted the attention of synthetic organic
chemists due to their interesting structures and promising phar-
maceutical applications.4 In recent years, much effort has been
made in order to synthesize stereochemically different and ring ex-
panded analogs of the above-mentioned aza-sugars because in
some cases, these analogs have shown increased glycosidase inhib-
itory and immunosuppressive activities.5 In this regard, Dhavale
and co-workers have successfully synthesized tetrahydroxylated
perhydroaza-azulenes 4a and 4b6 as well as a single seven-mem-
bered ring analog 55b (Fig. 1).
An efficient strategy for the preparation of pyrrolidines is
through a 5-exo-tet cyclization of a primary or secondary amine
onto an epoxide, which leads to the desired five-membered cyclic
amine bearing a hydroxyl group with known stereochemistry.6 In
this Letter, we report an accessible one-pot sequence for the
(+)-castanospermine (1)
(-)-swainsonine(2)
(+)-lentiginosine (3)
OH
H
N
OH
OH
H
OH
H
HO
OH
OH
OH
N
OH
N
HO
OH
HO
OH
4a
4b
5
Figure 1. Natural and synthetic polyhydroxylated alkaloids.
X
OH
O
Base
R
R
X
R
Et3B / O2
OH
7 (X=Br or I)
6 (X=Br or I)
8
Scheme 1. Radical-ionic strategy for the preparation of epoxides.
straightforward preparation of epoxides from either
a-iodoesters
or -iodonitriles and allylic alcohols based on a free-radical
a
Results and discussion
atom-transfer reaction followed by a base-promoted cyclization.
This sequence was applied to the synthesis of tetrahydroxylated
perhydroaza-azulene 4a.
Our strategy for the direct preparation of epoxides is depicted in
scheme 1. We reasoned that a Kharash-type reaction7 between an
alkyl halide (6), serving as the radical precursor, and an allyl alco-
hol would afford adduct 7, which upon treatment with a base,
would furnish the desired epoxide 8. The use of radical conditions
for the first step, which are neutral and compatible with other
⇑
Corresponding author. Tel.: +52 55 56224450; fax: +52 55 56162217.
E-mail address: acordero@unam.mx (A. Cordero-Vargas).
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.10.042

6900
E. Peralta-Hernández et al. / Tetrahedron Letters 52 (2011) 6899–6902
Table 1
Table 2
Optimization of reaction conditions
Different precursors and alcohols
O
O
Entry
1
Precursor
Alcohol
Product (Yield%)
A, B or C
OH
+
I
11a (49a, 78b, 80c)
EtO
9a
10a
EtO
O
O
O
9a
10a
11a
I
BnO
O
2
3
10a
Conditionsa
Initiator
Solvent(s)
Base
Yield (%)
BnO
11b (79%b, 61%c)
9b
A
B
C
Et3B
Et3B
Et3B
EtOH/H2O
CH2Cl2
(i)EtOH/H2O
(ii) CH2Cl2
K2CO3
DBU
DBU
49
78
80
OH
O
Ph
EtO
Ph
9a
O
10bd
11c (77%, 42%c)
a
Conditions A: Et3B (1.0 M solution in THF), O2 (trace), K2CO3 (2 equiv) in EtOH/
O
O
H2O at rt. Conditions B: Et3B (1.0 M solution in hexanes), O2 (trace), CH2Cl2; then
DBU (2 equiv). Conditions C: (i) Et3B (1.0 M solution in THF), O2 (trace), EtOH/H2O at
rt. (ii) DBU, CH2Cl2.
I
BnO
O
BnO
4
10a
Me
Me
9c
11d (15%b, 30%c)
O
NC
9d
I
NC
5
6
10a
reagents, would allow the use of an unprotected allylic alcohol and
the second step—ionic cyclization—to be done in a one-pot manner.
Table 1 shows the optimization of reaction conditions for this
transformation. We initially chose ethyl iodoacetate and allyl alco-
hol as model substrates. In the first attempt (conditions A), both
the radical initiator and the base were added since the beginning
of the reaction in an aqueous medium, leading to the isolation of
the desired product (11a), albeit in low yield (49%). Furthermore,
using dichloromethane as the solvent and DBU as the base (when
the radical reaction was completed), enhanced the yield to 80%
(conditions B). Finally, aqueous conditions for the radical step were
used again, but adding DBU as the base after switching the solvent
to CH2Cl2, affording epoxide 11a in 78% yield (conditions C).
Next, the scope of the procedure was studied for other sub-
strates (Table 2). Good yields were observed for primary stabilized
11e (83%b, 67%c)
OH
Ph
Ph
NC
9d
9d
O
10b
11f (71%b, 58%c)
HO Me
O
Me
O
O
O
O
7
NC
11g (47%b, 49%c)
10c
O
O
O
Br
O
O
8
10a
10a
10a
9e
11h (0%b, 30%c)
O
O
Et
I
Et
electrophilic radicals such as
-iodonitriles (9d, entries 5–7). In the case of substituted allylic
a-iodoesters 9a–b (entries 1–3) and
N
N
O
9
a
Et
Et
OH
alcohols (10b and 10c), nearly equimolar mixtures of cis:trans iso-
mers were observed in lower yields (entries 3, 6, and 7). The se-
quence also worked with a secondary electrophilic radical (9c,
entry 4) and for a bromolactone (9e, entry 8), albeit in modest
yields (30%). It is worth noting that for the latter substrate, the rad-
ical step only succeeded in ethanol–H2O, which is in agreement
with that reported by Fujimoto and co-workers, who proposed that
bromine atom transfer is better achieved in solvents with a high
dielectric constant.8 When N,N0-diethyliodoacetamide 9f was em-
ployed as a radical precursor (entry 9), the expected epoxide was
surprisingly not observed, but compound 11j, which is presumably
formed by opening of the initially formed epoxide by allyl alcohol.
9f
11i (45%b)
O
O
I
10
Cl
Cl
11j b,c,e
9g
I
Ph
11
12
10a
10a
S.M.b,c,f
9h
9i
S.M.b,c,f
I
a
b
c
Conditions A were used.
Conditions B were used.
Conditions C were used.
Compounds 10b–d were prepared by reaction of the corresponding aldehyde
with vinylmagnesium bromide according to known procedures.
For
a-iodoacetophenone 9g (entry 10), only the reduced product
11j was observed, probably due to the formation of the boron eno-
late prior to the radical addition.9 On the other hand, nucleophilic
radicals 9h and 9i failed to carry on the radical reaction (entries 11
and 12), which was somehow expected due to unfavorable polar
effects.10
d
e
Only 4-chloroacetophenone (reduced product) was detected after complete
consumption of the starting material and addition of 0.3 equiv of Et3B.
f
Only starting material was recovered after several hours of reaction.
The success of this sequential reaction when
a-iodonitriles
were used encouraged us to prepare 2-hydroxypirrolidines via a
5-exo-tet cyclization of 1-amino-5,6-epoxide. Typically, the route
for obtaining the mentioned epoxides involves a Claisen–Johnson
rearrangement of an allylic alcohol, followed by epoxidation, ester
reduction, transformation of the resulting alcohol into a leaving
group, substitution by the azide, and reduction of the latter to pro-
duce the primary amine which carries on the desired 5-exo-tet
cyclization.6 By using our protocol, the cyclization precursor would
be obtained in only 2 steps (one pot) from the corresponding allylic
alcohol (Scheme 2).
OH
O
O
nitrile
OH
R
R
R
reduction
5-exo-tet
R
NC
I
N
H
Et3B / O2
then DBU
CN
NH2
Scheme 2. Planned route for the synthesis of 2-hydroxypirrolidines.
ginal route described by Dhavale and co-workers (path A) involves
a 5 step linear sequence to prepare azide 13 from allylic alcohol 14.
The azide is then reduced and cyclized into the advanced interme-
diate 12. In our strategy, we planned to elaborate epoxy nitrile 15
Based on the above-mentioned strategy for the preparation of
2-hydroxypyrrolidines we proceeded to apply it to the synthesis
of perhydroaza-azulenes 4a and 4b. As shown in scheme 3, the ori-

E. Peralta-Hernández et al. / Tetrahedron Letters 52 (2011) 6899–6902
6901
A
OH
O
OH
(Dhavale et al.)
OH
OH
OH
H
N
5 steps
O
O
N
Cbz
OBn
OBn
14
O
O
deprotection-
reductive amination
O
O
N3
4a
12
13
nitrile reduction
-cyclization
B
(this paper)
O
OH
one pot
reaction
O
O
OBn
OBn
+
NC
9d
I
NC
O
O
O
O
radical addition
and epoxidation
15
14
Scheme 3. Described and proposed strategies for the synthesis of perhydroaza-azulene 4a.
O
as the key intermediate. The latter could be prepared in a more
O
straightforward manner from alcohol 14 and iodoacetonitrile
(9d) by applying our radical-ionic sequence and then transformed
into pirrolydine 12 by nitrile reduction and epoxide opening.
Thus, we started our synthesis by applying the radical addition-
NaBH4, NiCl2
CBz-Cl
O
NaH
N
N
O
DMF, 60 °C
OBn
O
15a
15b
OBn
MeOH, r.t.
O
Cbz NH
16a
O
O
O
O
epoxidation protocol to a mixture of D-xylofuranose derivative 14
17a
(85%)
(53%)
and iodoacetonitrile (9d). Surprisingly, with this particular sub-
strate, the radical step only proceeded when method C was used
(ethanol–H2O for the radical step and then DBU in DCM). Under
these conditions, a 6:4 diastereoisomeric mixture of 15a and 15b
was isolated in 68% yield (Scheme 4).
O
O
NaBH4, NiCl2
CBz-Cl
O
NaH
O
DMF, 60 °C
OBn
O
OBn
O
MeOH, r.t.
Cbz NH
16b
O
With epoxides 15a/15b in hand, we tested the reduction of the
nitrile moiety with different reagents and conditions. Catalytic
hydrogenation (H2/Pd–C) or strong reducing reagents (LiAlH4) only
lead to the recovery of the starting material or to a complex mix-
ture of products. After some experimentation, we found that mod-
ified Caddick conditions11 smoothly provided the Cbz carbamates
16a and 16b in 85% and 50% yields, respectively (Scheme 5). In or-
der to perform the desired 5-exo-tet cyclization, a number of meth-
ods, such as H2/Pd–C, BF3ꢁOEt2 or LDA were tested without success.
We finally found that when a mixture of 16a or 16b and NaH in
DMF was heated at 60 °C for 2–8 h, the expected cyclization is per-
formed but with the concomitant attack of the resulting alcoholate
over the carbamate, leading to the isolation of cyclic carbamates
17a (53%) and 17b (64%). 1H NMR spectra of 17a matched perfectly
with those reported previously.6 On the other hand, structure of
17b was unambiguously assigned by X-ray diffraction (Fig. 2).
The final steps of our approach are outlined in scheme 6. First,
cyclic carbamate 17a was opened by means of base treatment to
afford the corresponding aminoalcohol which was immediately
subjected to Cbz reprotection to afford 18a in 73% yield (2 steps).
Spectral data for compound 18a fully matched with those reported
by Dhavale and co-workers and it could be converted into perhy-
droaza-azulene 4a under the reported reaction conditions.6 Thus,
we have achieved a formal total synthesis of 4a. In the case of
17b, the same sequence of reactions was applied (carbamate
opening and Cbz protection), and 18b was isolated in 50% yield.
Compound 18b could be transformed into the new perhydroaza-
azulene (epimeric at C-6) by applying Dhavale’s conditions (as
for 18a).
O
17b
(50%)
(64%)
Scheme 5. Preparation of cyclic carbamates 17a and 17b.
Figure 2. ORTEP drawing of compound 17b.
NC
I
(9d)
1)
O
O
OH
Et3B, O2
EtO :H2O, r.t.
O
O
O
In conclusion, we have developed a useful free-radical atom-
transfer reaction followed by a base-promoted cyclization that
allows the direct obtention of epoxides from a-iodoesters or a-iod-
OBn
OBn
OBn
+
O
O
O
2) DBU, CH2Cl2
NC
NC
O
O
O
14
15a
15b
onitriles and allylic alcohols. The sequence can be performed in
either aqueous or organic solvents, without the use of protective
groups and in a one-pot fashion, which makes this sequence a very
68% (6:4)
Scheme 4. Synthesis of epoxides 15a and 15b.

6902
E. Peralta-Hernández et al. / Tetrahedron Letters 52 (2011) 6899–6902
1) NaOH
EtOH
OH
evaporated off, and the reaction purified by flash column
chromatography.
OH
OH
OH
OH
H
N
O
reflux
ref. 6
N
Cbz
OBn
17a
2) Cbz-Cl
NaHCO3
O
O
Acknowledgments
O
MeOH:H2O
18a
4a
(73%)
We wish to thank Instituto de Química (UNAM), CONACyT
(project number 80047) and DGAPA (project number I1200511)
for generous financial support. E. Peralta-Hernández and
O. Cortezano-Arellano also thank CONACyT for graduate scholar-
ships (grants number 23553 and 162040). We also thank Angeles
Peña and Isabel Chavez-Uribe for technical support (NMR).
1) NaOH
EtOH
OH
O
ref. 6
new analogs
reflux
N
Cbz
OBn
(epimeric at C-6)
17b
2) Cbz-Cl
NaHCO3
O
MeOH:H2O
18b
(50%)
Supplementary data
Scheme 6. Completion of the synthesis of perhydroaza-azulene 4a and of new
compound 18b.
Supplementary data (experimental procedures, spectroscopic
data for new compounds, and data for X-ray diffraction of com-
pound 16b) associated with this article can be found, in the online
version, at doi:10.1016/j.tetlet.2011.10.042.
useful tool in organic synthesis. We have also applied the above-
mentioned protocol to the formal synthesis of perhydroaza-azu-
lenes in a straightforward route. Efforts in order to develop a
stereoselective version of this protocol as well as the synthesis of
other biologically important alkaloids are currently in progress.
References and notes
1. (a) Hohenschutz, L. D.; Bell, E. A.; Jewess, P. J.; Leworthy, D. P.; Pryce, R. J.;
Arnold, E.; Clardy, J. Phytochemistry 1981, 20, 811; (b) Nash, R. J.; Fellows, L. E.;
Dring, J. V.; Stirton, C. H.; Carter, D.; Hegarty, M. P.; Bell, E. A. Phytochemistry
1988, 27, 1403.
Experimental Section
2. (a) Guennguerich, F. P.; DiMari, S. J.; Broquist, H. P. J. Am. Chem. Soc. 1973, 95,
2055; For a review on the synthesis of swainsonine and lentiginosine, see: (b)
One-pot protocol for the direct preparation of epoxides. (Meth-
od B): To a solution of the appropriate radical precursor (1 mmol)
and allylic alcohol (2 mmol) in dichloromethane (10 mL) at room
temperature was added triethylborane (0.1 mmol of a 1.0 M solu-
tion in hexanes). 2 mL of air was then bubbled into the solution
and the reaction mixture was allowed to stir at rt. Extra additions
of Et3B (0.1 mmol) and air (2 mL) were done every hour until com-
plete consumption of the starting material. DBU (1.5 mmol) was
then added and the reaction was stirred at rt for further 2 h. When
the reaction was completed, the solvent was evaporated under
reduced pressure and the residue was purified by flash
chromatography.
Nemr, A. E. Tetrahedron 2000, 56, 8579; For
a recent synthesis, see: (c)
Chooprayoon, S.; Kuhakarn, C.; Tuchinda, P.; Reutrakul, V.; Pohmaktor, M. Org.
Biomol. Chem. 2011, 9, 531.
3. (a) Pastuszak, I.; Molyneux, R. J.; James, L. F.; Elbein, A. D. Biochemistry 1990, 29,
1886; (b) Michael, J. P. Nat. Prod. Rep. 2001, 18, 520.
4. For reviews on polyhydroxylated alkaloids, see: (a) Sears, P.; Wong, C.-H.
Angew. Chem., Int. Ed. 1999, 38, 2300; (b) Chen, Y.; Vogel, P. J. Org. Chem. 1994,
59, 2487; (c) Michael, J. Nat. Prod. Rep. 2004, 21, 625.
5. (a) Burguess, K.; Henderson, I. Tetrahedron 1992, 48, 4045; (b) Markad, S. D.;
Karanjule, N. S.; Sharma, T.; Sabharwal, S. G.; Dhavale, D. D. Org. Biomol. Chem.
2006, 4, 3675; (c) Kumar, K. S. A.; Chaudhari, V. D.; Dhavale, D. D. Org. Biomol.
Chem. 2008, 6, 703; (d) Wang, N.; Zhang, L.-H.; Ye, X.-S. Org. Biomol. Chem. 2010,
8, 2639.
6. Markad, S. D.; Karanjule, N. S.; Sharma, T.; Sabharwal, S. G.; Puranik, V. G.;
Dhavale, D. D. Org. Biomol. Chem. 2006, 4, 2549.
Method C: To a solution of the starting material and allylic alco-
hol in ethanol/water (10 mL of a 3:1 mixture) was added Et3B
(0.1 mmol of a 1.0 M solution in THF). 2 mL of air was then bubbled
into the solution and the reaction mixture was allowed to stir at rt.
Extra additions of Et3B (0.1 mmol) and air (2 mL) were done every
hour until complete consumption of the starting material. Ethanol
was then evaporated and the reaction extracted with CH2Cl2. The
organic extracts were dried (Na2SO4), filtered, and concentrated
in vacuo. The residue was dissolved in 10 mL of dichloromethane
and 1.5 mmol of DBU was added to the reaction mixture. When
the starting material was completely consumed, the solvent was
7. (a) Byers, J. In Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-
VCH: Weinheim, 2001; pp 72–89; Atom transfer radical reactions can also be
catalyzed by Cu(I). For a review, see: (b) Eckenhoff, W. T.; Pintauer, T. Cat. Rev.
2010, 52, 1.
8. Yorimitsu, H.; Nakamura, T.; Shinokubo, H.; Oshima, K.; Omoto, K.; Fujimoto, H.
J. Am. Chem. Soc. 2000, 122, 11041.
9. (a) Nozaki, K.; Oshima, K.; Utimoto, K. Bull. Chem. Soc. Jpn. 1991, 64, 403; (b)
Horiuchi, Y.; Taniguchi, M.; Oshima, K.; Utimoto, K. Tetrahedron Lett. 1995, 36,
5353; (c) Chandrasekhar, S.; Narsihmulu, C.; Reddy, N.; Reddy, S. Tetrahedron
Lett. 2003, 44, 2583.
10. (a) Giese, B. Angew. Chem., Int. Ed. 1983, 22, 753; (b) Tedder, J. M.; Walton, J. C.
Acc. Chem. Res. 1976, 9, 183.
11. Caddick, S.; Judd, D. B.; Lewis, A. K. K.; Reich, M. T.; Williams, R. V. Tetrahedron
2003, 59, 5417.